Investigating mechanisms can help determine which medications show the most promise for alcoholism treatment. Naltrexone (NAL) results in improved treatment outcomes but there are no studies of the effects of 100 mg dose on mechanisms such as urges or responses to drinking. Ondansetron (OND) shows considerable promise with little information about its mechanisms. The combination of OND with NAL reduces drinking but has never been tested against each medication alone. Certain individual difference variables have shown promise for moderating the effects of some medications. The objective of this research is to evaluate the value of OND (0.5 mg/day), NAL (100 mg/day) and both together on several mechanisms along with several hypothesized moderating variables. The specific aims are to investigate: (1)whether these doses of OND, NAL or the combination results in reduced cue-elicited responses to alcohol cues among abstinent alcoholics in treatment (Study 1); (2) whether OND and/or NAL results in reduced pleasurable effects and craving after alcohol ingestion among non-treatment-seeking alcoholics (Study 2); (3) whether OND and/or NAL results in less alcohol consumption in a bar and in the natural environment (Study 2); (4) the extent to which the effects in the above aims are moderated by early vs. late-onset alcoholism, family history of alcoholism, or presence of the 7-repeat allele of the D4 dopamine receptor. Two studies will be conducted. The first will investigate the effects of 7 days on each medication on reactivity to alcohol cues among recently abstinent alcoholics in treatment, using a 2 X 2 (NAL vs. placebo; OND vs. placebo) between-groups randomized placebo-controlled design. The second will investigate the effects of 7 days on each medication on reactions to alcohol ingestion in a bar-lab and on amount consumed both in the bar-lab and the home environment among non-treatment-seeking heavy drinkers, using a 2 X 2 within-subject placebo-controlled cross-over design. Results will provide evidence for the relative promise of these medications separately and in combination on measures relevant to relapse and severity of drinking. Evidence will also be provided about subpopulations more likely to benefit from each medications.